Most of the 20 million individuals with chronic kidney disease (CKD) die before commencing dialysis. One out of every 5 dialysis patients dies each yr, a mortality worse than most cancers. Although cardiovascular (CV) disease appears to be the most common cause of death in CKD, CV risk factors such as hyper- cholesterolemia, hypertension, and obesity are paradoxically associated with better survival in hemodialysis patients. Emerging data indicate the existence of the CV paradoxes even in earlier CKD stages and in chronic heart failure (CHF) patients. There are also paradoxical relationships by race and ethnicity. African Americans, who have lower life expectancy and higher CKD prevalence than Caucasians, survive significantly longer when undergoing dialysis. Hispanic Americans who paradoxically have lower than expected rates of chronic diseases have exceptionally higher CKD prevalence. Better understanding of the mechanisms for these paradoxes may lead to new strategies to improve survival in these populations and have meaningful clinical and public health implications. We hypothesize that: (A) these paradoxes may evolve progressively over the natural course of CKD as a result of the time discrepancy between the 2 competing risk factors, i.e., overnutrition (long-term killer) vs. undernutrition (short-term killer);(B) a process of "survival selection" over the course of CKD or CHF progression may lead to populations with different pathophysiology;and (C) similarities between CHF and CKD may help disclose their common pathophysiologic mechanisms. In the spirit of the PA-06-163 we propose to study the Southern California Kaiser Permanente database with >1 million CKD patients over 10 yrs, the DaVita national database with -200,000 hemodialysis patients over 8 yrs, and UCLA CHF database with >2,800 CHF patients over 20 yrs. Using novel epidemiologic methods including time-dependent survival and nested structural modeling and incidence density risk-set matching, we will examine the associations between predictors and survival in graded strata of traditional risk factors across racial/ethnic groups and CKD stages. The following questions will be examined: (1) Are there clinical constellations leading to the "African American Dialysis Paradox" beyond statistical confounding and survival selection bias? (2) Is the "CKD Paradox within the Hispanic Paradox" a consequence of diabetes or other clinical conditions? (3) Are the "CV Risk Factor Paradoxes" due to mere residual confounding or can they imply non-traditional pathophysiologic mechanisms? (4) Can the "CHF Paradox" and its similarities with and differences from the CKD paradoxes contribute to better understanding of the CKD paradoxes? Relevance to Public Health: Examining the CKD and CHF paradoxes may lead to developing new treatment targets beyond traditional (Framingham) CV risk factors to improve poor outcomes in these 2 patient populations and maybe in geriatric populations or those with chronic disease states and similar paradoxes